PROJECT SUMMARY The long-range goal of this work is to predict cyclophosphamide (CY) pharmacokinetics using a mechanistic- based marker of plasma endogenous metabolomics compounds (EMC) and covariates identified with pharmacokinetic (PK) modeling (i.e., EMC/PK-model). The prodrug CY has an increasing role in allogeneic hematopoietic cell transplant (alloHCT), which cures patients of their underlying disease by replacing their hematopoietic system with that of a healthy donor. CY has complex metabolic pathway with substantive (up to 16-fold) variability in the area under the plasma concentration-time curves (AUCs) of CY and its metabolites with current mg/kg dosing. Until recently, one of the most significant limitations to optimizing CY dosing was the difficulty in quantifying plasma concentrations of 4hydroxycyclophosphamide (4HCY). An understanding of 4HCY pharmacokinetics is critical because 4HCY in turn forms phosphoramide mustard, which covalently cross-links DNA. Because phosphoramide mustard does not cross cell membranes easily, the transport of its precursor, 4HCY, into the cell is a key step of CY?s cytotoxic activity. Ex-vivo, 4HCY suppresses the growth of bone marrow progenitor, CFU-GM, natural killer, and T-cells in a concentration-dependent manner. However, conducting pharmacokinetic/ pharmacodynamic studies to evaluate if plasma 4HCY AUC is associated with clinical outcomes is too resource?intensive because of multi-center studies are needed to achieve sufficient power. Thus, we seek to develop a mechanistic EMC/PK-model that predicts the ratio of 4HCY/CY AUC and can be evaluated as a mechanistic-based biomarker in subsequent clinical trials. In Aim 1, we will identify the EMC that predicts the ratio of 4HCY/CY AUC. Using targeted metabolomics, we discovered nine EMC and one pathway statistically associated with the ratio of 4HCY/CY AUC. These results guided our choice of evaluating two different targeted metabolomics assays in a larger (n=279 patients) training dataset. We will subsequently conduct separate in vitro (human hepatocyte) and in vivo (prospective cohort of n=100 patients) validation. In Aim 2, we will describe CY and its metabolites pharmacokinetics using empirical and physiological models. Our laboratory pioneered 4HCY quantitation. We were the first to personalize CY doses, specifically by using population pharmacokinetic (popPK)-guided CY dosing. Age is the only covariate in this popPK model. Using the remainder of our unique PK dataset, we seek to identify new covariates in a more heterogeneous population of various diseases, ages, and combination CY regimens. For those covariates identified by popPK modeling, we will then create a physiologically-based pharmacokinetic model (PBPK) to simulate the impact of these covariates on the ratio of 4HCY/CY AUC. These results will then be used with those of Aim 1 to create a EMC/PK-model that predicts the ratio of 4HCY/CY AUC. In addition, the results of Aim 2 will impact patient care by being used for clinical practice guidelines for CY dosing in alloHCT.